includes/SiteStats.php

   1 <?php
   2 /**
   3  * Accessors and mutators for the site-wide statistics.
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  */
  22
  23 /**
  24  * Static accessor class for site_stats and related things
  25  */
  26 class SiteStats {
  27         /** @var bool|ResultWrapper */
  28         private static $row;
  29
  30         /** @var bool */
  31         private static $loaded = false;
  32
  33         /** @var int */
  34         private static $jobs;
  35
  36         /** @var int[] */
  37         private static $pageCount = array();
  38
  39         /** @var int[] */
  40         private static $groupMemberCounts = array();
  41
  42         static function recache() {
  43                 self::load( true );
  44         }
  45
  46         /**
  47          * @param bool $recache
  48          */
  49         static function load( $recache = false ) {
  50                 if ( self::$loaded && !$recache ) {
  51                         return;
  52                 }
  53
  54                 self::$row = self::loadAndLazyInit();
  55
  56                 # This code is somewhat schema-agnostic, because I'm changing it in a minor release -- TS
  57                 if ( !isset( self::$row->ss_total_pages ) && self::$row->ss_total_pages == -1 ) {
  58                         # Update schema
  59                         $u = new SiteStatsUpdate( 0, 0, 0 );
  60                         $u->doUpdate();
  61                         self::$row = self::doLoad( wfGetDB( DB_SLAVE ) );
  62                 }
  63
  64                 self::$loaded = true;
  65         }
  66
  67         /**
  68          * @return bool|ResultWrapper
  69          */
  70         static function loadAndLazyInit() {
  71                 wfDebug( __METHOD__ . ": reading site_stats from slave\n" );
  72                 $row = self::doLoad( wfGetDB( DB_SLAVE ) );
  73
  74                 if ( !self::isSane( $row ) ) {
  75                         // Might have just been initialized during this request? Underflow?
  76                         wfDebug( __METHOD__ . ": site_stats damaged or missing on slave\n" );
  77                         $row = self::doLoad( wfGetDB( DB_MASTER ) );
  78                 }
  79
  80                 if ( !self::isSane( $row ) ) {
  81                         // Normally the site_stats table is initialized at install time.
  82                         // Some manual construction scenarios may leave the table empty or
  83                         // broken, however, for instance when importing from a dump into a
  84                         // clean schema with mwdumper.
  85                         wfDebug( __METHOD__ . ": initializing damaged or missing site_stats\n" );
  86
  87                         SiteStatsInit::doAllAndCommit( wfGetDB( DB_SLAVE ) );
  88
  89                         $row = self::doLoad( wfGetDB( DB_MASTER ) );
  90                 }
  91
  92                 if ( !self::isSane( $row ) ) {
  93                         wfDebug( __METHOD__ . ": site_stats persistently nonsensical o_O\n" );
  94                 }
  95                 return $row;
  96         }
  97
  98         /**
  99          * @param DatabaseBase $db
 100          * @return bool|ResultWrapper
 101          */
 102         static function doLoad( $db ) {
 103                 return $db->selectRow( 'site_stats', array(
 104                                 'ss_row_id',
 105                                 'ss_total_views',
 106                                 'ss_total_edits',
 107                                 'ss_good_articles',
 108                                 'ss_total_pages',
 109                                 'ss_users',
 110                                 'ss_active_users',
 111                                 'ss_images',
 112                         ), false, __METHOD__ );
 113         }
 114
 115         /**
 116          * @return int
 117          */
 118         static function views() {
 119                 self::load();
 120                 return self::$row->ss_total_views;
 121         }
 122
 123         /**
 124          * @return int
 125          */
 126         static function edits() {
 127                 self::load();
 128                 return self::$row->ss_total_edits;
 129         }
 130
 131         /**
 132          * @return int
 133          */
 134         static function articles() {
 135                 self::load();
 136                 return self::$row->ss_good_articles;
 137         }
 138
 139         /**
 140          * @return int
 141          */
 142         static function pages() {
 143                 self::load();
 144                 return self::$row->ss_total_pages;
 145         }
 146
 147         /**
 148          * @return int
 149          */
 150         static function users() {
 151                 self::load();
 152                 return self::$row->ss_users;
 153         }
 154
 155         /**
 156          * @return int
 157          */
 158         static function activeUsers() {
 159                 self::load();
 160                 return self::$row->ss_active_users;
 161         }
 162
 163         /**
 164          * @return int
 165          */
 166         static function images() {
 167                 self::load();
 168                 return self::$row->ss_images;
 169         }
 170
 171         /**
 172          * Find the number of users in a given user group.
 173          * @param string $group name of group
 174          * @return int
 175          */
 176         static function numberingroup( $group ) {
 177                 if ( !isset( self::$groupMemberCounts[$group] ) ) {
 178                         global $wgMemc;
 179                         $key = wfMemcKey( 'SiteStats', 'groupcounts', $group );
 180                         $hit = $wgMemc->get( $key );
 181                         if ( !$hit ) {
 182                                 $dbr = wfGetDB( DB_SLAVE );
 183                                 $hit = $dbr->selectField(
 184                                         'user_groups',
 185                                         'COUNT(*)',
 186                                         array( 'ug_group' => $group ),
 187                                         __METHOD__
 188                                 );
 189                                 $wgMemc->set( $key, $hit, 3600 );
 190                         }
 191                         self::$groupMemberCounts[$group] = $hit;
 192                 }
 193                 return self::$groupMemberCounts[$group];
 194         }
 195
 196         /**
 197          * @return int
 198          */
 199         static function jobs() {
 200                 if ( !isset( self::$jobs ) ) {
 201                         $dbr = wfGetDB( DB_SLAVE );
 202                         self::$jobs = array_sum( JobQueueGroup::singleton()->getQueueSizes() );
 203                         /**
 204                          * Zero rows still do single row read for row that doesn't exist,
 205                          * but people are annoyed by that
 206                          */
 207                         if ( self::$jobs == 1 ) {
 208                                 self::$jobs = 0;
 209                         }
 210                 }
 211                 return self::$jobs;
 212         }
 213
 214         /**
 215          * @param int $ns
 216          *
 217          * @return int
 218          */
 219         static function pagesInNs( $ns ) {
 220                 wfProfileIn( __METHOD__ );
 221                 if ( !isset( self::$pageCount[$ns] ) ) {
 222                         $dbr = wfGetDB( DB_SLAVE );
 223                         self::$pageCount[$ns] = (int)$dbr->selectField(
 224                                 'page',
 225                                 'COUNT(*)',
 226                                 array( 'page_namespace' => $ns ),
 227                                 __METHOD__
 228                         );
 229                 }
 230                 wfProfileOut( __METHOD__ );
 231                 return self::$pageCount[$ns];
 232         }
 233
 234         /**
 235          * Is the provided row of site stats sane, or should it be regenerated?
 236          *
 237          * Checks only fields which are filled by SiteStatsInit::refresh.
 238          *
 239          * @param bool|object $row
 240          *
 241          * @return bool
 242          */
 243         private static function isSane( $row ) {
 244                 if ( $row === false
 245                         || $row->ss_total_pages < $row->ss_good_articles
 246                         || $row->ss_total_edits < $row->ss_total_pages
 247                 ) {
 248                         return false;
 249                 }
 250                 // Now check for underflow/overflow
 251                 foreach ( array(
 252                         'ss_total_views',
 253                         'ss_total_edits',
 254                         'ss_good_articles',
 255                         'ss_total_pages',
 256                         'ss_users',
 257                         'ss_images',
 258                 ) as $member ) {
 259                         if ( $row->$member > 2000000000 || $row->$member < 0 ) {
 260                                 return false;
 261                         }
 262                 }
 263                 return true;
 264         }
 265 }
 266
 267 /**
 268  * Class designed for counting of stats.
 269  */
 270 class SiteStatsInit {
 271
 272         // Database connection
 273         private $db;
 274
 275         // Various stats
 276         private $mEdits = null, $mArticles = null, $mPages = null;
 277         private $mUsers = null, $mViews = null, $mFiles = null;
 278
 279         /**
 280          * Constructor
 281          * @param bool|DatabaseBase $database
 282          * - Boolean: whether to use the master DB
 283          * - DatabaseBase: database connection to use
 284          */
 285         public function __construct( $database = false ) {
 286                 if ( $database instanceof DatabaseBase ) {
 287                         $this->db = $database;
 288                 } else {
 289                         $this->db = wfGetDB( $database ? DB_MASTER : DB_SLAVE );
 290                 }
 291         }
 292
 293         /**
 294          * Count the total number of edits
 295          * @return int
 296          */
 297         public function edits() {
 298                 $this->mEdits = $this->db->selectField( 'revision', 'COUNT(*)', '', __METHOD__ );
 299                 $this->mEdits += $this->db->selectField( 'archive', 'COUNT(*)', '', __METHOD__ );
 300                 return $this->mEdits;
 301         }
 302
 303         /**
 304          * Count pages in article space(s)
 305          * @return int
 306          */
 307         public function articles() {
 308                 global $wgArticleCountMethod;
 309
 310                 $tables = array( 'page' );
 311                 $conds = array(
 312                         'page_namespace' => MWNamespace::getContentNamespaces(),
 313                         'page_is_redirect' => 0,
 314                 );
 315
 316                 if ( $wgArticleCountMethod == 'link' ) {
 317                         $tables[] = 'pagelinks';
 318                         $conds[] = 'pl_from=page_id';
 319                 } elseif ( $wgArticleCountMethod == 'comma' ) {
 320                         // To make a correct check for this, we would need, for each page,
 321                         // to load the text, maybe uncompress it, maybe decode it and then
 322                         // check if there's one comma.
 323                         // But one thing we are sure is that if the page is empty, it can't
 324                         // contain a comma :)
 325                         $conds[] = 'page_len > 0';
 326                 }
 327
 328                 $this->mArticles = $this->db->selectField( $tables, 'COUNT(DISTINCT page_id)',
 329                         $conds, __METHOD__ );
 330                 return $this->mArticles;
 331         }
 332
 333         /**
 334          * Count total pages
 335          * @return int
 336          */
 337         public function pages() {
 338                 $this->mPages = $this->db->selectField( 'page', 'COUNT(*)', '', __METHOD__ );
 339                 return $this->mPages;
 340         }
 341
 342         /**
 343          * Count total users
 344          * @return int
 345          */
 346         public function users() {
 347                 $this->mUsers = $this->db->selectField( 'user', 'COUNT(*)', '', __METHOD__ );
 348                 return $this->mUsers;
 349         }
 350
 351         /**
 352          * Count views
 353          * @return int
 354          */
 355         public function views() {
 356                 $this->mViews = $this->db->selectField( 'page', 'SUM(page_counter)', '', __METHOD__ );
 357                 return $this->mViews;
 358         }
 359
 360         /**
 361          * Count total files
 362          * @return int
 363          */
 364         public function files() {
 365                 $this->mFiles = $this->db->selectField( 'image', 'COUNT(*)', '', __METHOD__ );
 366                 return $this->mFiles;
 367         }
 368
 369         /**
 370          * Do all updates and commit them. More or less a replacement
 371          * for the original initStats, but without output.
 372          *
 373          * @param DatabaseBase|bool $database
 374          * - Boolean: whether to use the master DB
 375          * - DatabaseBase: database connection to use
 376          * @param array $options of options, may contain the following values
 377          * - views Boolean: when true, do not update the number of page views (default: true)
 378          * - activeUsers Boolean: whether to update the number of active users (default: false)
 379          */
 380         public static function doAllAndCommit( $database, array $options = array() ) {
 381                 $options += array( 'update' => false, 'views' => true, 'activeUsers' => false );
 382
 383                 // Grab the object and count everything
 384                 $counter = new SiteStatsInit( $database );
 385
 386                 $counter->edits();
 387                 $counter->articles();
 388                 $counter->pages();
 389                 $counter->users();
 390                 $counter->files();
 391
 392                 // Only do views if we don't want to not count them
 393                 if ( $options['views'] ) {
 394                         $counter->views();
 395                 }
 396
 397                 $counter->refresh();
 398
 399                 // Count active users if need be
 400                 if ( $options['activeUsers'] ) {
 401                         SiteStatsUpdate::cacheUpdate( wfGetDB( DB_MASTER ) );
 402                 }
 403         }
 404
 405         /**
 406          * Refresh site_stats. If you want ss_total_views to be updated, be sure to
 407          * call views() first.
 408          */
 409         public function refresh() {
 410                 $values = array(
 411                         'ss_row_id' => 1,
 412                         'ss_total_edits' => ( $this->mEdits === null ? $this->edits() : $this->mEdits ),
 413                         'ss_good_articles' => ( $this->mArticles === null ? $this->articles() : $this->mArticles ),
 414                         'ss_total_pages' => ( $this->mPages === null ? $this->pages() : $this->mPages ),
 415                         'ss_users' => ( $this->mUsers === null ? $this->users() : $this->mUsers ),
 416                         'ss_images' => ( $this->mFiles === null ? $this->files() : $this->mFiles ),
 417                 ) + (
 418                         $this->mViews ? array( 'ss_total_views' => $this->mViews ) : array()
 419                 );
 420
 421                 $dbw = wfGetDB( DB_MASTER );
 422                 $dbw->upsert( 'site_stats', $values, array( 'ss_row_id' ), $values, __METHOD__ );
 423         }
 424 }